This paper proposes a puzzle about an agent’s beliefs that arises in certain cases where a person suffering from depression makes sincere assertions of negative self-evaluations because of her depression whilst also judging that her overall evidence supports the negation of her negative self-evaluations. I argue that the assertions of negative self-evaluations and the agent’s judgments about overall evidence each exhibit attributes that are characteristic of belief. I argue that the puzzle arises in cases where the agent in question is aware of each of her beliefs and is rational enough to prefer, not only avoiding conflicting beliefs, but also to prefer that her beliefs be based on relevant overall evidence, as opposed to depressive thought patterns. I consider a variety of theories that may map on to my case, and I discuss challenges or objections to each theory as it applies to the case. I think there are reasons to favor a theory that rejects the claim that the agent straightforwardly believes that her negative self-evaluations are true, so, after ruling out some unsatisfying accounts, I spend the first major portion of the paper discussing views that take this approach. The latter portion of the paper is spent discussing theories that put pressure on other premises or assumptions of the puzzle. My intention is to discuss the relationship between depression, belief, judgment, and rationality in a somewhat narrow context. I do not intend to speak to the relationship between belief, judgment, and rationality as it applies to depression in general. I leave it up to further inquiry to discuss depression more generally as it relates to the issues raised in this paper, and I leave it up to further inquiry to determine whether my proposed puzzle or its potential solutions have novel or interesting therapeutic implications.